Eye Gaze Based Hands-Free Navigation in Virtual Environments

Abstract

Eye tracking has shown great potential in supporting interaction tasks in VR, such as selection, manipulation and navigation. Previous studies have shown that eye gaze based interaction performs poorly for selection and manipulation tasks due to the low accuracy of eye gaze data. Navigation, however, requires less accuracy than selection and manipulation. In this thesis, a hands-free virtual navigation technique that utilizes eye gaze data is proposed. The technique allows users to use eye gaze for virtual navigation while leaving their hands free for other more important interaction tasks. The proposed system requires minimal physical interaction; therefore, it could also be beneficial to VR users with severe motor disabilities. A user study with 11 participants compared the proposed technique to head tilt based navigation using an obstacle avoidance navigation task. The results of the study indicate that eye gaze and head tilt have comparable performance, which is promising in comparison to previous studies. This shows that the proposed eye gaze based virtual navigation technique may be a viable navigation technique in VR.